Source of the Oxygen in the C-O-P Linkage of the Acyl Phosphate in Transport Adenosine Triphosphatases*

SUMMARY The source of the oxygen in the C-O-P linkage of the acyl phosphate in the Naf,Kf-ATPase of porcine kidney and the Ca2+,Mg2+-ATPase of rabbit muscle sarcoplasmic reticulum has been assessed. Formation of the Na+,K+-ATPase acyl phosphate in presence of ouabain and [‘*O]Pi results in the C-O-P bridge oxygen remaining nonisotopic. Thus, the mode of entry of Pi into the acyl phosphate is by attack of a carboxylate oxygen on the phosphorus atom and displacement of a hydroxyl group. When the acyl the the C-O-P phosphate cleaved oxygen the acyl to an subsequent attack of


SUMMARY
The source of the oxygen in the C-O-P linkage of the acyl phosphate in the Naf,Kf-ATPase of porcine kidney and the Ca2+,Mg2+-ATPase of rabbit muscle sarcoplasmic reticulum has been assessed.
Formation of the Na+,K+-ATPase acyl phosphate in presence of ouabain and ['*O]Pi results in the C-O-P bridge oxygen remaining nonisotopic. Thus, the mode of entry of Pi into the acyl phosphate is by attack of a carboxylate oxygen on the phosphorus atom and displacement of a hydroxyl group.
When the acyl phosphate of the Ca2+,Mg2f-ATPase or the Na+, Kf-ATPase is formed in the presence of H180H, the C-O-P bridge oxygen also remains nonisotopic.
These results show that the acyl phosphate is cleaved by water oxygen attack on the phosphorus atom and rules out any acyl transfer to an enzyme group and subsequent hydrolysis by attack of water oxygen on the acyl carbon atom.
The phosphoryl oxygens of the Naf,K+-ATPase phosphoenzyme at 15" derived either from ["O]Pi in the presence of ouabain or from ATP in the presence of HnOH were observed to undergo exchange with medium water, not accounted for by exchange of Pi oxygens with HOH and incorporation of the Pi into the protein-bound acylphosphate.
The cleavage of ATP coupled to ion transport by microsomal preparations in presence of Na+ and K+ or by muscle sarcoplasmic reticulum preparations in presence of Ca2+ and Mg2+ is known to proceed through intermediate phosphorylation of a protein carboxyl group (l-5).
Three possibilities warrant consideration for the cleavage of this acyl phosphate intermediate of the transport ATPases.
One is by direct attack of a water oxygen on the phosphoryl phosphorus atom to form Pi. A second is by attack of water oxygen on the acyl carbon atom, dis-placing Pi. A third is by attack of another protein group on the acyl carbon with displacement of Pi, essentially an acyl transfer, followed by subsequent hydrolysis of the acyl derivative.
All three modes of cleavage will, in the usual assay conditions where many molecules of ATP are cleaved per active site, give rise to Pi which derives at Ieast 1 oxygen from water.1 The first mode of cleavage can be distinguished from the other two by measurement of the source of oxygen in the first Pi formed per active site in ATP cleavage, or by determination of the source of the bridge oxygen of the C-O-P linkage in the acyl phosphate after continued ATP hydrolysis.
Cleavage resulting from water attack on the phosphoryl phosphorus atom will, in subsequent reaction cycles, give an acyl phosphate with the bridge oxygen derived from the carboxyl group.
Conversely, cleavage by attack of water oxygen or another group on the acyl carbon atom will, in subsequent reaction cycles, give an acyl phosphate with the bridge oxygen derived from water.
In addition to the phosphorylation by ATP, the transport ATPases can, under appropriate conditions (668), be phosphorylated from Pi in the absence of ATP.
Phosphorolysis of an acyl linkage or displacement by attack of a carboxylate oxygen on the phosphorus atom of Pi could occur. Again, the mode of phosphorylation can be ascertained by measurement of the source of the C-O-P bridge oxygen in the phosphorylated enzyme .
It is the purpose of this paper to present findings on the source of the C-O-P bridge oxygen in the microsomal and sarcoplasmic reticulum ATPases phosphorylated from either Pi or ATP. prepared as described previously (8). Endogenous 1'; ranged from 0.8 to 3 pmoles per 100 mg of protein.
Sarcoplasmic reticulum vesicles were prepared from rabbit skeletal muscle as described by Kanazawa and Boyer (7).

Determination of 180 Content of Bridge Oxygen of Membrane
Acyl Phosphate-Relatively large scale preparation of the respective phosphoenzymes was carried out as described in the text. For P-O bond cleavage, the acid-washed, protein-bound acyl phosphate was heated in 0.1 N formic acid, pH 3.8, for 15 min at 98". For C-O bond cleavage, heating for 15 min at 100" was performed in 0.1 N NaOH.
Following hydrolysis in the presence of carrier Pi, the reaction mixture was deproteinized with perchloric acid, adjusted to pH 8. and extracted with an equal volume of buffered aqueous phenol (88% phenol, pH 8) to remove perchloric acid-soluble protein and lipid components (10) that otherwise interfere with quantitative extraction of the Pi. The phenol lower layer was subsequently extracted with an equal volume of water, and the supernatanta were combined.
The resulting solution was cooled to near 0" and centrifuged, and the supernatant was made 1 x in HCl and extracted with 4 volumes of isobutyl alcohol-benzene (1: 1, v/v) to remove residual phenol. The organic phase was discarded, and the aqueous layer was made 0.012 M in ammonium molybdate.
Phosphate was separated and I*0 was determined as described previously (8).
was prepared essentially as described by Glynn and Chappell (11) or Post and Sen (12).
Protein concentrations were estimated by the Lowry procedure with bovine serum albumin as a standard.
+P-Labeled ATP small release of 32P represented largely acyl phosphate hydrolysis at 4". That the residual 32P present with the precipitated protein was covalently bound in this and other experiments was to be expected from previous reports of others (l-6).
Further evidence for this is given by the observation that the 32P remained with the protein when it was solubilized by 6 M guanidine HCl at pH 2 and reprecipitated by 0.4 M perchloric acid. In addition, conditions for hydrolytic release of 32Pi from the precipitated protein are those expected to cleave acyl phosphate bonds.
Separate internal standards employed ["O]Pi addition to denatured protein samples to correct the observed I80 content for small apparent losses of I80 accompanying phosphate isolation, purification, and analysis. Such internal standards were found to be indispensable with the amounts of phosphoprotein conveniently available.
Corrections for internal standards ranged from 10 to 12%.
Results given in Table II show that the observed 180 content of Pi resulting from P-O cleavage was identical, within experi- gen-The determination of the source of the bridge oxygen is based upon known cleavage patterns of acetyl phosphate in alkali or acid, resulting in C-O and P-O cleavage, respectively. The Pi isolated from an acyl phosphate cleavage in an appropriate acid medium should contain the 3 phosphoryl oxygens plus an oxygen from medium water.
With alkaline cleavage, the 3 phosphoryl oxygens plus the C-O-P bridge oxygen would be present in the Pi. Thus, comparison of the I80 content of the Pi after C-O and P-O cleavage allows calculation of the I80 content of the bridge oxygen. ,4 restriction in the design of experiments, however, is the relatively small amount of acyl phosphate protein available compared to the Pi sample size and 180 enrichment required for accurate analysis.
To test the conditions anticipated to give P-O and C-O cleavage (13, 14), appropriate hydrolyses were carried out with acetyl phosphate in the presence of H180H.
Results obtained with hydrolysis in dilute NaOH or in formate buffer are shown in Table I. They establish clear preferential C-O or P-O cleavage with acetyl phosphate under the conditions uped.
For the experiments with acetyl phosphate (Table I), a temperature of 75" was used. The lability of protein-bound acyl phosphate might differ somewhat; hence a temperature of 100" was used to help assure complete cleavage. Ouabain was added to promote formation of higher amounts of the phosphoenzyme (6). In related studies, phosphoenzyme formation was shown to be complete within 2 s after ouabain addition.
The reaction was terminated at 40 s by the addition of 200 ml of 0.4 M perchloric acid containing 20 mM Pi. Phosphoprotein was isolated as described in the text and divided into three portions.
Cleavage in presence of 4.9 pmoles of carrier Pi, Pi isolation, and I80 analysis with corrections for internal standards were made as described in the text. Results are averages of the triplicate determinations f S.D. b Corrected for internal standards as described in the text. 3), 300 mg of electroplax Na+,K+-ATPase (specific activity = 5.2 rmoles per min per mg of protein) in a total volume of 8.0 ml of water containing 9.64 atom y. 180. Also, for Experiment 3, 20 mM KC1 was present. The reactions were initiated by addition of ATP and quenched after 15 s by the rapid addition of 300 11 of 5.7 M trichloroacetic acid containing 0.1 M Pi. Cleavage in presence of 4.9 pmoles of carrier Pi, Pi isolation, and '80 analyses were made as described in the text.
The reaction mixture contained 5 InM [Y-~~P]ATP, 10 mM MgCL, 0.5 mM CaCl2, 100 KCl, 100 mM Tris-HCl (pH 7.0), and 98.4 mg of sarcoplasmic reticulum ATPase in a total volume of 5.0 ml of Hr*OH (10.34 atom s excess) at 15". The reaction was initiated upon addition of ATP and quenched after 10 s by the rapid addition of 299 hl of 5.7 M trichloroacetic acid. Inorganic phosphate was isolated following hydrolysis of the acid-washed, proteinbound acyl phosphate under the conditions selective for P-O or C-O cleavage as described in the text. b Corrected for internal standards as described in the text. c Twice as much protein and ATP were used. mental error, with the isO content of Pi resulting from C-O cleavage, showing that the bridge oxygen was not 'SO-enriched. The expected 180 atom per cent excess for the sample with C-O cleavage if Pi furnished the bridge oxygen would be one-third greater than that for P-O cleavage, or 0.071 atom $Zo excess. Clearly, the upper limit of the observed atom per cent excess was well below the expected atom per cent excess if Pi furnished the bridge oxygen. These data demonstrate that acyl phosphate formation occurs by a direct displacement of a phosphoryl oxygen by the carboxylate oxygen.
Also of interest is the observation that the isO content of the phosphoryl oxygens of the acyl phosphate was about 30% lower than that expected on the basis of the 180 content of [i80]Pi used. This indicates a phosphate-water exchange at the level of the phosphoenzyme intermediate.
This has been confirmed in other experiments.
With longer reaction times in presence of ouabain extensive exchange of the phosphoryl group oxygens of the acyl phosphate-ouabain complex occur. Source of Bridge Oxygen in Naf,K+-ATPase Phosphorylated by ATP in Presence of H1gOH-For these experiments, the enzyme was allowed to cleave ATP in Hi80H, and the possible incorporation of water oxygens into the acyl phosphate bridge oxygen was measured. Table III presents data on the '*O content of the Pi resulting from P-O or C-O cleavage of separate, identical samples of the phosphoenzyme formed from ATP in the presence of H180H. Controls and internal standards similar to those presented in the text for Table II were performed. Observed atom per cent excesses of '80 in the Pi isolated are given to indicate possible OOl could be accounted for by that derived from the phosphoryl oxygens. Thus, no measurable exchange of the carboxyl group oxygens with water occurred during catalysis.
The observed atom per cent excess of igO reported in Table III when corrected for carrier Pi addition shows that a relatively large amount of water oxygens was incorporated into the phosphoryl oxygens of the acyl phosphate.
Such exchange amounted to an average of about 1 oxygen per phosphoryl group.
One possibility for this IsO incorporation from water is that a portion of the phosphoenzyme present was formed from medium Pi containing ISO derived by ATP hydrolysis or by Pi * HOH exchange (8). However, control similar experiments with presence of 32Pi indicated that less than 5% of the phosphoenzyme present during ATP hydrolysis arose from medium Pi. This is quite insufficient to account for the observed 180 in the enzyme phosphoryl group.
Exchange of Hr80H with phosphate oxygens has thus occurred prior to formation of medium Pi. Such exchange as observed with myosin ATPase is referred to as intermediate exchange (15).

Source of Bridge
Oxygen in Ca2+, Mgzf-ATPase Phosphorylated by ATP in Presence of HigOH-Phosphorylation of sarcoplasmic reticulum vesicles was carried out in H180H with ATP as substrate as described in Table IV. The observed incorporation of 180 into the Pi formed from either P-O or C-O cleavage was less than 0.001 atom y. excess. The results show that water oxygen was not incorporated into either the C-O-P bridge oxygen or the phosphoryl oxygens of the acyl phosphate during ATP cleavage. Measurements of the steady state level of the phosphoryl enzyme and Ca2f-dependent Pi liberation indicate that the phosphory enzyme turned over at least 20 times during the reaction period. DISCUSSION The results demonstrate that with the microsomal and the sarcoplasmic reticulum ATPases, the acyl phosphate intermediate is cleaved by water as shown in Equation 1. Incorporation of water oxygen is indicated by the filled oxygen atoms. sources of error.
In our laboratory, samples of the size used give 180 values measurable to about ~kO.002 atom To excess. The experiment is thus approaching the sensitivity of the method.
Enzyme-C-0--POz= + (1) The results show no experimentally significant differences between the '80 content of the Pi derived by P-O or C-O cleavage.
HOH + enzyme-COOH + HOPOI' Within the experimental error of about 5 to lo'%, all 180 present The data rule out the possibility of formation of any internal acyl-X derivatives in the membrane as a mode of utilization of the acyl phosphate for active transport.
Also, they rule out attack of water on the acyl carbon in preference to the phosphoryl phosphorus atom.
The results on the phosphorylation of the microsomal ATPase from Pi in presence of ouabain demonstrate that this occurs by direct displacement of a phosphate oxygen by the enzyme carboxylate, essentially a reversal of Equation 1. The active site in some manner may increase the nucleophilicity of the carboxylate oxygen, increase the susceptibility of the phosphorus oxygen to nucleophilic attack, or favor departure of the OH--leaving group.
The formation of a substantial equilibrium amount of an acyl phosphate from Pi without phosphorolysis of a pre-existing covalent bond points to conformational or other factors favoring existence of the phosphorylated form. Such a capacity, promoted by ouabain binding, may be germane to suggested conformational coupling for oxidative phosphorylation (16). Mention should be made that the experimental limitations are such that the IsO analyses by themselves do not establish that the enzyme-COOH group is the sole source of the C-O-P bridge oxygen.
They do establish that it is the major source and, within experimental limitations, could be the only source. The high probability that the enzymic reactions involved in the phosphorylation are completely specific as to position of bond cleavage and formation allow the conclusions made herein.
The demonstration in these studies of a preferential exchange of water oxygens with Pi released from the phosphorylated Na+,K+-ATPase is quite analogous to the "intermediate" exchange observed in hydrolysis by myosin and actomyosin ATPases (15). Previous studies from this laboratory showed considerable exchange capacity of the Na+, K+-ATPase but only with Pi oxygens of the medium (8). The essential difference appears to be that the earlier studies were conducted at 37" whereas the present results were obtained at 15". Temperaturedependent changes in mechanism of the Na+,K+-ATPase have been proposed previously by Kanazawa et al. (17). Conformational changes in the 14 to 20" region are known to occur which have been invoked to explain biphasic Arrhenius plots of Na+, Kf-ATPase activity (18). Rate-limiting steps in ATP hydrolysis at the lower temperature may include release of Pi from the enzyme form as depicted by Step 2 of Equation 2.